Loading...
  OR  Zero-K Name:    Password:   

Returning observation

287 posts, 6090 views
Post comment
Filter:    Player:  
Page of 15 (287 records)
sort
quote:
What if you had 2 factory tokens to start with? (This could be a mod-option)

That was my wish thinking too, when many facs was the meta. But then factory prices were raised from 600 to 1000 instead (now 800). I think the idea is to force you more towards sticking with your soft faction rather than having you only use your favorite units. Having more factories available does kind of put the work put into factory lineups to waste.
+3 / -0

4 years ago
I'm reluctant to have rooms based on rank rather than culture. Fighting the small team vs lobsterpot culture wars has at least taught me that any option that mechanically coercing certain game styles whether by accident or design will face quite a lot of hostility. Streaming players to particular lobbies would have that effect and we don't have the numbers to get away with it. Additionally, I'm happy to play with novices who share my desire to play to win, but simply lack the experience to do it. Helping to develop them is a rewarding side-benefit.

That said, coloured leagues without numbers seems like a low hanging fruit.

I'm also reluctant to try to fix an issue that lies in the head of the player with a balance change in game. The slower the game starts and the harder it is to engineer an advantage through a raid that had not been prepared for, the easier it will be to slip into turtle mode. That's already the lowest energy state a team game (especially a lobsterpot) falls into and it's also the most frustrating state because the only individual who is likely to break the deadlock is Firepluk who has been rushing a Zenith from the start. Everyone else is stuck fighting an attrional war for scraps of reclaim that Pluk's Athenas will probably suck up anyway since he's devoting all his attention to finding metal rather than fighting. Having Pluk be your admittedly erratic win button isn't really a healthy place for a team game.

People might find the lack of real agency in a lobsterpot reassuring but it's false reassurance, and however little control you have over the outcome, the mighty WHR machine will still judge you. That's why I'm increasingly convinced of the merit of the no-elo default. If you don't actually have much control over the outcome, why get blamed for losing?

That said, since early radar is usually a no-brainer anyway, saving 55 metal and putting it on the commander would be one little bit less near mandatory micro to worry about. Perhaps it could even be tuned so a stay at home engineer commander has a slightly larger range, and the go-getting recon has slightly less range balancing the increased mobility. Simply seeing trouble coming is always the best defence. Over-committing to the offence should be punished as the tactical error it is.

I also don't see why wind farms should chain explode. Just to be clear I presume this refers specifically to them blowing each other up on destruction rather than simply being very easy to kill by running a glaive down the line? Making a chain of e is a tactical choice in the hands of the player. If it is overly attractive vs alternatives despite the risk of catastrophic failure, maybe it does need a bit of a tweak. I don't imagine a real wind turbine would explode on destruction of its own accord so why can't vicious war-fighting robots work out that packing a stick of dynamite in their turbine doesn't improve efficiency?
+4 / -0
quote:
weren´t you the one that taught me to use badgers against moderators

I've been laughing in the face of everyone who suggested that for the past year or so.

quote:
don't imagine a real wind turbine would explode on destruction of its own accord so why can't vicious war-fighting robots work out that packing a stick of dynamite in their turbine doesn't improve efficiency?

IIRC the stick of dynamite is supposed to damage the raider, and thus limit its potential total impact. Windgens killed pre-range-buff fleas well enough. I've seen plenty of too greedy scythes die to accumulated damage from windsplosions. Perhaps this would work better if they did more damage in higher AoE and visually this could be represented by blade shrapnel rather than actual dynamite fireball.

quote:
What if you had 2 factory tokens to start with? (This could be a mod-option)

I don't think fac variety needs that much of a tip on the scale to make it improve. Incrementally better maps, incrementally better balance. Revenant torpedos and general flexibility/integration improvements like that. Maybe expanding the list of roles each factory has to have (with antisub, but maybe also terrain negotiation, etc. Lobster absolutely does terrain negotiation on spider/jump level for amphs, so why can't jumps or even rovers attack sea? Well, dominatrix now can have subs, i applaud that)

quote:
That said, since early radar is usually a no-brainer anyway, saving 55 metal and putting it on the commander would be one little bit less near mandatory micro to worry about. Perhaps it could even be tuned so a stay at home engineer commander has a slightly larger range, and the go-getting recon has slightly less range balancing the increased mobility. Simply seeing trouble coming is always the best defence. Over-committing to the offence should be punished as the tactical error it is.

In the adjacent discussion about Kodachi, the two games that i could investigate - ones that ROrankForever casted - had UltraGojira's opponents fail to make radar and get cooked for that. So default-on radar (on non-recon chassis for interestingness? instead of the overpowered jump?) sounds quite interesting.
+4 / -0
Ah that makes it clearer and a bit more sense Squig. I'm with you now. Small piece of input back from me rn (short on time); Recon comm IS the meta in high tier 1v1. If you put the radar on the engineer and came up with some other perks for the other two commanders maybe they'd return to viability too.

+1 for rev torps.
+2 / -1
4 years ago
quote:
Additionally, I'm happy to play with novices who share my desire to play to win, but simply lack the experience to do it.


This. So much this. Ranks and elo tell you nothing about how motivated a player is to win. For example, I'd take FRrankNitekiller or RUrankProKidAshKa in my team over a truckload of supergiants or neutrons; even though their individual skill is on the low side, these are people that can be communicated with and they play in a team. Elo is completely blind to all of this, so the "fair teams based on elo" in pots are sometimes extremely demotivating.
+0 / -0
It seems to me that many agree at least at on considering ideas of built-in radar for (not recon - good call yet ironic) coms and color tier ladder/league with detailed ladder positions revealed seasonally.

What can we do to make it more real than the usual forum talk? Can any admin make poll of community votes for/against those changes with detailed arguments for and against such changes?

I know that design decisions are to be made by maintainers and are not really the democration process, but it could be. It would certainly make community feel more empowered to contribute if our forum discussions can result in some possitive changes. Usually it's just a bunch of hot air and all discussions fade away with effort put into bringing arguments and suggestions and 9/10 nothing else happens.
+2 / -1


4 years ago
Gf's already paying attention to the thread Zen so that's a big help where these things are concerned, if my past experience with him is anything to go by. I've seen at least 3-4 ideas mentioned to him during my "intimate" spell with the game be implemented - so the dude definitely listens to this stuff. Obviously the scope of changes is always varying and some things are too disruptive. I think that's where you do have to pay attention to the democratic will of the community to find the correct path.
+0 / -1
4 years ago
Think it would be good to have a poll, it is hard to know how representative are the posts in the thread.

Check for example second poll at https://zero-k.info/Poll/UserVotes/5295 - 65% of people want overall performance and not "recent" - which could be associated with the leagues idea. Of course people do change opinion and preference, the league idea was not in the pool and don't even remember when that pool was made.
+2 / -0
4 years ago
Katastrophe: I often open the game with raiders and never attack the enemy's base. Their speed makes them useful for defending my expansion while hindering the enemy's. Riots are too expensive and slow for early map control.

On the subject of radars, how about a durable variant? Radar is another thing that tends to get shredded by artillery shots, raids or whatever and you take a while to notice. The "radar visualizer" thing could also be easier to understand and on by default.

On the subject of coms, the Recon chassis does feel OP atm, specially if you have no interest in com upgrades. The higher movement speed and jump are so bloody useful in the early game, it blows anything else the other coms have to offer out of the water. Whereas the engineer chassis is pointless if you don't upgrade.

Zenfur: To be honest, I don't want design decisions to become a democratic process. "Many cooks spoil the food" is a strong phenomenon in game development. I just participate in these discussions in the hopes of that a dev may read them and feel inspired to make a positive change in the game. After all even someone who develops a game doesn't see every angle that players might see.
+1 / -0


4 years ago
quote:
Vehicles of any sort in particular are awfully handled by the game without player input because all pathfinding and tactical AI systems mostly consider them to be able to turn in place, which they can't. Some of them also have hilariously bad combinations of stats, like Scorchers whose slowly turning turrets mean you not only have to micromanage your raiders, but also subsystems of those raiders.

For me, it doesn't feel bad to lose to good plays. It feels absolutely awful to lose to these various malfunctions.

THIS! SO MUCH THIS!
I HATE HATE HATE the fact that some units have slow turret turn speed, there is NO need for it and from a user perspective it just means those units will be buggy a lot of the time.

That said, for me this preference has nothing to do with why I don't play 1v1 too much, it's just a general reduction in my enjoyment.
+2 / -0
BRrankManored:

ok, then i will ask from another point of view:

how long does 1 glaive need to destroy:

1 Solar (closing included)
1 Factory
1 Windgen
1 Mex

yes, it´s only one glaive, but it´s about the difference in time, not the total amount.


solars take ages, factories take ages, windgens die fast, mexes die fast.


i want to adress 2 things here:

- if you are worried about your energy-structures, go for solars, especially on easy to acces terrain.

- mexes already have an implemented buffer, reclaiming them gives you back a good amount of their cost. Thats why it´s better to raid energy instead of mexes. e-stalling means you can´t use the reclaim at once, so you don´t have that buffer. Even a larger amount of glaives has trouble killing solars without standing still for some time, wich is pretty detrimental to then normally.

To put it bluntly:
I think the raider-problem is exaggerated.

1. Sure, you will lose 1 or 2 mexes sometimes, but this is the "natural" flow of the game and does not hurt as much as long as you rebuild fast enough. Expansion is a fluid process, nothing that reaches a stable point and then never changes again. If a raid gets ALL your mexes, yes thats pretty detrimental, but a lot of things have to have been gone wrong beforehand that that happens.

2. To protect my builders Riots are sufficient because they are not slower than builders usually. And if i don´t need to protect my base from enemy-raiders, why would i have to build some of them myself at all? ? Pork is way better locally.
+1 / -0
4 years ago
But if you go solars while your enemy goes windmills, the enemy will generally have the advantage, which creates a pressure for going windmills. "There are safer options" isn't a solution when there is an economic pressure to take the unsafe options.

About 2, because all your raiders can be in one spot, the enemy porc can't. And the only way to effectively stop them from creating local numeric superiorities is with raiders of your own. Off course the importance of this is map-dependent, in a 1v1 with a reasonably sized map with lots of mex spots, its pretty important.
+0 / -0


4 years ago
From what I can see, these are the main themes touched on so far.
  • Expansion speed and health.
  • The ranking system.
  • Unit behaviour.
  • Factory choice/matchups.

quote:
Seriously, one glaive or flea should not be able to destroy an entire energy farm so easily just because it used it's insane speed to insert itself into a gap that the opponent could not reasonably afford in resources nor micro-time to defend.

quote:
Mexes are also part of that equation of economic frailty, given the exponential effect that lost mexes have in your economy thanks to overdrive. And its easy to miss the fact that a mex was destroyed in the mid-late game, which makes you further paranoid.

quote:
If expanding was slower, and expansions were more resilient, having such keyed reflexes might not be so important.

Compare to dawn of war for example, where listening posts took a good 20 seconds for a force to focus down, and taking down main base structures took forever. Dawn of war was appreciably non-stressful to play.

I think durability and the economy are the easiest and potentially most fruitful things to tinker with.

The question of whether mexes are too cheap has been lingering over development for many years. A mex pays for itself in about 35s. Raids are nowhere near that frequent, so you should be grabbing mexes as fast as you can. The main expense is cosntructors, but they can take territory with metal-efficient turrets. Overall, the window for exploiting the economic dip caused by constructing a mex is very small. Therefore, raids are focused on attrition rather than exploiting windows. I think mexes could easily be quite a bit healthier and costlier.

Health is interesting because it increases the window for reacting and nerfs high-alpha units, which themselves are a bit frustrating. As a concrete example consider ZK with doubled health for everything. After a bit of rebalancing, what does it look like and is it a reasonable game? Would it be correct to say that low-health raiders encourage defenses and counter-raids, while high health could encourage reinforcing an area under attack?

Thinking about health and alpha got me thinking about Lotus and Picket. These two units could be a microcosm for the ways that an RTS can be particularly stressful. Pickets often shoot before they are spotted and quick reflexes are required to avoid losing a raider pointlessly. Attacking a cluster of Pickets is very much all-or-nothing. Pickets can be frustrating on the defending end when one of their shots fails to hit (often due to a moving target going behind terrain) or when overkill prevention malfunctions. Lotus suffers from none of these problems, and I think the game is better for the change that relegated Picket to 100m cost (up from 80m).

Also on raiders (they are important for 1v1), what do people feel about the retreat 'range bonus'? This is the phenomena in which only one of two identical fighting units can fire, depending on their speed and spacing. I noticed a few streamers get frustrated at it closer to the Steam release, perhaps some people here feel the same. It is core to raider interactions and makes for some nice defenders advantage. I don't know if or how this could be removed as it is simple a result of the physics of the game. Changing the physics is difficult and has even jankier side effects. Something drastic would need to be done, such as making units slower when firing (a micro nightmare).

Doubling health across the board is not without costs. One of my criticisms of some of the more recent RTS games is that it feels like their weapons have little impact. The battles are fairly static and silly looking when a relatively small unit takes 100 bullets to kill. Reducing the impact of alpha would also soften the balance.

quote:
Whatever constant tinkering has happened to the system up unto this point has totally undermined the purpose of it and left us with inflated numbers

quote:
That seems like a great idea to decrease rank anxiety. Instead having ladder available at all times, make it an event to show the ranking order over some period of time - like every 4 months - together with recap of all most achieved and biggest awards. Automate that even.

I think seasons for competitive MM is a very good idea, as well as a fake rating that approximately reflects your 'real' skill but which moves in ways that are more fun than the current system. I don't know how to do any of this short of taking on responsibility for infrastructure development. Basically, CHrankAdminDeinFreund took on responsibility for ratings with the driving motivation of applying some fancy tech to make them better predictors of games. Unfortunately, predicting games accurately led to concessions on the side of usability - remember when your rating would 'randomly' go down on a loss? Over time, a lot of work was put in to smooth out the system to make it more appropriate for human consumption, but it is still a system with accuracy instead of human-readability at its core.

I think most people want a simple, understandable, system with predictable results. Random losses are terrifying, so not being able to predict how much you may lose of a loss surely contributes to stress. Seasons would be a massive hit to game predictability, but they would allow people to come back and "reset" without the pressure to maintain their previous rating.

quote:
I'm also reluctant to try to fix an issue that lies in the head of the player with a balance change in game.
Briefly, hoping that people change is completely backwards game design. Sure, you can design a game that is "not for" the majority of people, but then you'd better be happy with doing that. The premise of this thread is what might be done if we're not happy with that.

quote:
I think a simple change which would make the game a lot less stressing without needing any actual gameplay changes, would be making units less propense to suicide.

quote:
I HATE HATE HATE the fact that some units have slow turret turn speed, there is NO need for it and from a user perspective it just means those units will be buggy a lot of the time.

I am not affected by the first issue (or you could say I was highly affected) because I set my units to perpetually hold position in as many games as will allow it. I have noticed this issue, and even tried hold position as the new install default for a time. Hold position is not a good general solution because then units sit there and look stupid while they die to something just outside range, I could see that new players were better off with Maneuver. I don't really know why people would want their units to move automatically so I don't have a way to start making good idle unit AI.

Slow turret turn speed is slowly being eliminated just based on the way patch notes have gone. I have not seen too many problems due to unit turn speed. I worry that if vehicles turned significantly faster then they would essentially become bots, losing a large part of their identity. However, there is a lot of room between their current behaviour and becoming bots. Some reduction in turn radius could be applied.

quote:
There's so many maps where you HAVE to go X factory to be competitive - and that's a stressor; not a precursor for fun and expression.

quote:
What if you had 2 factory tokens to start with? (This could be a mod-option)

I don't think a feeling of having to go factory X can be solved by itself. If you are already stressed by the game then you are more likely to stick to what is safe than experiment and extend the meta. It is also hard to experiment while competitively stressed, which will feed back into the feeling that only factory X is viable. I agree that factory balance could be better, or there may even be better ways to select your 'faction'.

Starting with two factory tokens just sounds like it will make the game harder. I don't see what would prevent the meta settling in the way that @Sparkles brings up and, lets be honest, one of the factories will be air (making the game even harder). I also think that 1v1 is the only game mode suffering from factory choice problems. In teamgames, even down to 2v2, factory choice distinguishes the roles of players in a team. There are many combinations of factories and the simplicity makes it easy for experienced players to know what types of things they can do for a team. That ZK can be played in large team games owes a lot to the current factory system.

I don't see what can be done short of a full redesign of the factory system. Most designs seem like they would run into the same problems, and potentially break teamgames. The current system of attempting to balance the factories while picking maps that don't overly favour one factory has worked in the past.

====================

If @Sparkles or anyone else wants to test some drastic changes I'm happy to put them on a test host. From this post, the drastic thing I would test is:
  • Double health.
  • 100 cost mex.
  • Increase turn rates of low turn rate things.
I have no idea if it works, but it seems best to test extremes.
+5 / -0

4 years ago
I will gladly play any test host with changes. However double health does present a large task of potential rebalancing. Some unit interactions will for sure change drastically.
+0 / -0
4 years ago
I feel the "retreat range bonus" is frustrating mostly because its unintuitive, as well as because units can easily be baited into it, creating another major source of micro. I think it would be best if it were removed or if unit AI could handle it better, but dunno what a good way to go about doing that would be.

I feel like the current ranks aren't doing much to create balanced games, and might actually be doing worse than random selection would do. Some players have their skill level differ greatly from their actual rank, but off course the autobalancer doesn't know that. For example, a high level player that for some reason is playing badly will be put up with low level players repeatedly, sometimes leading to a situation where everyone already knows which team will win in advance. That is frustrating for everyone involved.

I feel that an ideal idle unit AI would be more defensive than offensive, aka, it focuses on keeping itself alive and secondarily holding the position you left them in, without pushing beyond that.

The problem is how can the unit know where the "frontline" is so that it doesn't cross it? Maybe players could have "zoning" tools to help guide unit ai? As in, you draw a circle where units are allowed to maneuever, and idle units inside won't leave the circle unless ordered (maybe let them enter a contiguous circle though). Or you could draw no-go zones instead where units won't go unless explicitly ordered. Dunno how viable implementing something like this would be.

Regarding the factory system, I have personally never liked it. I feel that mixing unit movement types and "mini factions" into the same system is inherently over-complicated and unintuitive. For example, why do hovers have lance and heavy vehicles don't have their own mobile tachyon? Because faction themes. However, faction themes are something only a veteran understands. A newbie has no idea shield bots have shit artillery until they need it and its not there. Zero-K's factory system practically demands that a player memorize the entire unit roster to understand what tools they can get where, which is crazy. It also creates a certain rock-paper-scissors element to factory match-ups.

I am not certain of how Zero-K's factory system could be redesigned given that its largely a done game, however, if I had to, I'd start by breaking the strict 1:1 relationship between units and factories, aka, allow certain units to be built by multiple factories if it would make sense, and redesign the factories to be more about playstyles than movement modes.
+1 / -0


4 years ago
I would already consider most factories to be about playstyles rather than movement types. The factories which are really about movement types, such as planes, spider and amph, arguably demonstrate the playstyle suited to that movement type.

Perhaps I should spell out the mechanics that create the "retreat range bonus" just to demonstrate that it is the simplest result of the game physics, and that there are no intuitive solutions. Here are some facts about unit physics:
  • Weapons have range.
  • Weapons have a flight time, after which many weapons dissipate or lose propulsion power. Most weapons have flight time just beyond their range.
  • Units only fire at enemies within their range.
  • Units predict the movement of enemies, and only fire at enemies that will be within their range when the projectile reaches the target.
If unit A in range of and moving towards unit B, then unit B can fire because its target will still be in range. If unit B is moving away from unit A, then unit A will not fire unless its target will be in range (current range!) when the projectile reaches it.

What alternatives are there?
  • Units don't check whether their target will be in range - Units will fire when their projectile flight time is insufficient to reach the target.
  • Units don't check whether their target will be in range, and projectile flight times are boosted - I would expect this to quickly dissolve the physiscy verisimilitude of ZK. What happens when a target is faster than a projectile? Infinite range.
  • Make all projectiles so fast that the effect is negligible - Erodes the balance levers and identity of ZK. I believe this to be the solution used in SupCom. There must be some solution because I've talked to some people about this and they swear SupCom does not have "retreat range bonus".
  • Make projectiles inherit unit velocity (aka Galilean relativity) - Opens a whole new can of worms. Units gain effective range on whatever they are moving towards, and lose range against things they are moving away from. Prepare for artillery and skirmishers running up to fire like a bowler.
+0 / -0
4 years ago
quote:
Make projectiles inherit unit velocity (aka Galilean relativity) - Opens a whole new can of worms. Units gain effective range on whatever they are moving towards, and lose range against things they are moving away from. Prepare for artillery and skirmishers running up to fire like a bowler.

This sounds the most in the spirit of the game. Could be compensated for artillery by making all artillery stop when fire. Then for skirmishers the projectile speed could be much higher than the skirmisher speed then the effect will be lower.

Does this require engine modifications?
+2 / -0


4 years ago
I quite like the "retreat bonus" of raiders giving an advantage to defenders. It prevents snowballing quite as badly in the early game.
+3 / -0


4 years ago
I don't like the idea of balance changes, ZK balance is tight. Perhaps just experiment with hiking mex cost + HP?

I like the idea of having abstract divisions instead of a visible player ladder. Perhaps just have a visible number 1 2 and 3 spot.
+2 / -0

4 years ago
I support seasons as soon as they make sense. One should probably play at least twenty games per season to get any kind of reasonable rank. If we don't want the number of players who fulfill that requirement to shrink below the 50 ladder slots that means 6-12 months per season. For inactive players like myself that would mean I probably always hang around 1500 rating as I don't play enough games to rank up before the next rating reset. So all of the less active pros of which we have a lot would turn the newbie regions into a smurf fest. Imagine drone comes back for a handful of games at the end of a season and pummels the top 10 while rated as a newbie.

So all I see coming out of this is a less accurate ladder, nightmare matchmaking for newbies and a free smurf card given out once every rating reset. Do we really need that free smurf card to keep players happy?
+1 / -0
Page of 15 (287 records)